We have a monitoring system on our Windows 7 Ultimate ( 64 bit, SP 1 ) that pings 1000+ IPs every 30 seconds to record latency and "uptime". I have 3 IPs so far that work on every other connection/computer I try. I have verified firewall ( hardware ) is doing everything properly. I am able to ping the same IPs from another server we have at the same location. I have tried pinging with the software firewall turned off and the anti-virus program off and no effect. I have rebooted twice ( before and after the most recent updated for windows ), both times I have been able to ping those IPs for probably about 10+ minutes and then I get host unreachable from then on. I started using Wireshark to packet sniff and I see the ping packet is getting a reply, which is very odd. Obviously the host is reachable because the ping packet reply came back in. So the question is what is causing the ping ICMP packet to report as unreachable. Anyone have any ideas? I changed the IP to xx.xx in the example below as well as the reply packet. asked 05 May ‘11, 15:21  wolfjlupus edited 05 May ‘11, 15:46  SYN-bit ♦♦ |
One Answer:
The ICMP ID's are not the same. So the PING program will not match the response to the request. It looks like a device between 192.168.2.112 and 64.105.xx.xx is messing up the ICMP ID's. Is there a NAT device in place between those hosts? Maybe it's tables are not large enough to translate that many ICMP packets. answered 05 May '11, 15:50 SYN-bit ♦♦ |
Thank you Synbit, further packet sniffing at the firewall level shows it is the Sonicwall doing improper NATing even though it's set up properly. Once I removed the outbound NAT policy for it's specific IP it started working. The Sonicwall probably needs to be factory reset as it's pretty old and has been running for a long time.